The essence of bootstrapping: resampling with replacement to estimate a sampling distribution and its statistics of interest.

bootstrapping.R

library(tidyverse)
library(gganimate)

# sample of 100 data points
tibble(x = rexp(1e2)) -> d

# 20 bootstrap samples, you likely want several orders of magnitude more.
tibble(boot_samples = seq_len(2e1)) %>%
  mutate(data = map(boot_samples, ~ d %>% sample_frac(1, replace = TRUE))) %>%
  unnest(cols = c(data)) %>%
  ggplot(aes(x = x, group = boot_samples)) +
  stat_ecdf(size = 0.25) +
  xlab("Time to event") +
  ggtitle("Bootstrap samples from one sample of an\n
          Exponential distribution r.v. with rate parameter 1") +
  transition_reveal(boot_samples) -> anim

animate(anim, fps = 10, duration = 2)